Agentic AI Accelerates SETI Search at Allen Telescope Array
Breakthrough Listen and the SETI Institute deploy NVIDIA-powered AI agents to steer the Allen Telescope Array and process 2.4 terabits per second live.
Summary
On October 30, 2026, Breakthrough Listen and the SETI Institute brought agentic AI into daily searches at Northern California’s Allen Telescope Array. From one natural language request, agents can steer antennas, calibrate and focus the instrument, correct data errors, apply AI models and plot results. NVIDIA Holoscan supplies live sensor data, helping compress work that took humans weeks or years into hours. The approach could search for advanced technology, including hypothetical AI created by extraterrestrial civilizations billions of years older than humanity.
DAQIRI, a Holoscan library, transfers terabits per second directly into GPU memory for processing in flight. The SETI Institute’s CyberEther framework performs signal processing with NVIDIA CUDA-X libraries cuFFT and cuBLAS. Nexus, a web orchestrator and agent harness, tracks instrument state, compute resources and pipelines while researchers monitor packet loss, inspect live spectrograms and rewrite processing blocks through prompts. NVIDIA RTX PRO 6000 Blackwell GPUs handle 2.4 terabits per second, with inference running on premises beside the telescope.
The pipeline achieved the first real-time AI detection of a pulsar in the Crab Nebula. The team will next evaluate open models including NVIDIA Nemotron for rapid, local processing and potential offline observatories on the moon’s far side. Andrew Siemion, the SETI Institute’s Bernard M. Oliver chair for SETI and Breakthrough Listen’s principal investigator, frames possible contact between human and extraterrestrial AI as a translation breakthrough rather than a threat.
Positives
- 2.4 terabits per second can be processed on NVIDIA RTX PRO 6000 Blackwell GPUs beside the telescope.
- Weeks or years of human work can be compressed into hours by agents controlling and correcting observations.
- The first real-time AI detection of a pulsar in the Crab Nebula validates the pipeline.
- One natural language request can steer antennas, calibrate instruments, repair errors and plot results.
- NVIDIA Nemotron could support rapid local processing and future offline observatories on the moon’s far side.
Risks & concerns
- Terabits per second of telescope data create a substantial real-time processing and storage challenge.
- Packet loss and data errors still require continuous monitoring and correction during observations.
- Extraterrestrial civilizations and alien-built AI remain hypothetical, leaving the system’s ultimate target unconfirmed.
- NVIDIA Nemotron and other open models are still being evaluated for local and offline use.
- A far-side lunar observatory remains a future possibility rather than an operating deployment.